xfs
[Top] [All Lists]

Re: What is needed for a stable 2.4 based system?

To: Jan-Frode Myklebust <janfrode@xxxxxxxxxxx>
Subject: Re: What is needed for a stable 2.4 based system?
From: Christian Guggenberger <Christian.Guggenberger@xxxxxxxxxxxxxxxxxxxxxxxx>
Date: Thu, 13 Feb 2003 02:41:51 +0100
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <20030211201605.A16099@ii.uib.no>; from janfrode@parallab.no on Tue, Feb 11, 2003 at 20:16:05 +0100
References: <200302101000.22190.krienke@uni-koblenz.de> <Pine.LNX.4.44.0302100707170.20309-100000@stout.americas.sgi.com> <20030211201605.A16099@ii.uib.no>
Sender: linux-xfs-bounce@xxxxxxxxxxx
Hi,
On 11.02.2003   20:16 Jan-Frode Myklebust wrote:
On Mon, Feb 10, 2003 at 07:08:32AM -0600, Eric Sandeen wrote:
> For starters, the patch you downloaded is a development snapshot, so if
> it's stable, you're just lucky - it has not undergone extensive testing.

Is it really that bad? I'm in the process of setting up a server with
~2.5 TB of XFS storage, and was just about to go for the linux-2.4-xfs
cvs-tree. Plain 2.4.19 woun't do for me, since there's a hard lockup
bug in the tg3 driver that wasn't fixed before 2.4.20rc3:


I'm using xfs-cvs since last August on an NFS Server with 4.3 TB with almost no problems at all. We had some NFS server problems, but these got fixed only some days after that bug got public (Bug #186, I think). Other problems were related to Intel's e1000 driver, which got better with 2.4.20. For our use, xfs-cvs _is_ very stable. But, of course, it's always good to test with your Envrionment, before leaving for production use...
Besides this, you always could ask for patches for the tg3 (Jeff Garzik, I think, is the man who cares about that drivers on lkml) against 2.4.19!


Christian


<Prev in Thread] Current Thread [Next in Thread>